Abstract
This study aimed to enhance the structural performance of a drift joint with a slotted-in steel plate using laminated timber made from small- to medium-diameter timbers. An analysis was conducted to compare and evaluate the structural performance and seismic behavior of laminated timber portal frames reinforced with fiber-reinforced plastic (FRP) laminates, such as glass fiber cloth (GFC), glass fiber reinforced plastic sheet (GS), and carbon-reinforced plastic sheet (CS), under low amplitude cyclic lateral loads, in comparison to unreinforced portal frames (UR). The reinforced portal frame, strengthened by FRP laminates, exhibited higher resistance to strength and stiffness strength degradation due to cyclic loading-induced degradation at low drift angles compared to the unreinforced frame affected by heterogeneous timber. The reinforced portal frame, particularly the GFC and GS, exhibited energy dissipation rates 32% and 17% higher than UR, respectively, making them the most effective in vibration damping. FRP laminates mitigated fiber direction cleavage from drift pin holes. The average maximum moment and average yield moment of the reinforced portal frame increased by 18% and 21%, respectively, due to the alleviation of joint failure. Finally, CS significantly affected the maximum moment, while GFC significantly influenced the yield moment.
Download PDF
Full Article
Behavior of Laminated Timber Portal Frames Reinforced with Fiber-reinforced Plastic Laminates under Cyclic Lateral Load
Yo-Jin Song,a Seong-Yeob Baek,b Hyun-Woo Kim,b and Soon-Il Hong b,*
This study aimed to enhance the structural performance of a drift joint with a slotted-in steel plate using laminated timber made from small- to medium-diameter timbers. An analysis was conducted to compare and evaluate the structural performance and seismic behavior of laminated timber portal frames reinforced with fiber-reinforced plastic (FRP) laminates, such as glass fiber cloth (GFC), glass fiber reinforced plastic sheet (GS), and carbon-reinforced plastic sheet (CS), under low amplitude cyclic lateral loads, in comparison to unreinforced portal frames (UR). The reinforced portal frame, strengthened by FRP laminates, exhibited higher resistance to strength and stiffness strength degradation due to cyclic loading-induced degradation at low drift angles compared to the unreinforced frame affected by heterogeneous timber. The reinforced portal frame, particularly the GFC and GS, exhibited energy dissipation rates 32% and 17% higher than UR, respectively, making them the most effective in vibration damping. FRP laminates mitigated fiber direction cleavage from drift pin holes. The average maximum moment and average yield moment of the reinforced portal frame increased by 18% and 21%, respectively, due to the alleviation of joint failure. Finally, CS significantly affected the maximum moment, while GFC significantly influenced the yield moment.
DOI: 10.15376/biores.19.4.8739-8757
Keywords: Laminated timber; Portal frame; Joint; Lateral load; Cyclic test; Reinforcement; FRP
Contact information: a: Institute of Forest Science, Kangwon National University, Chun-Cheon 200-701 Republic of Korea; b: Department of Forest Biomaterials Engineering, Kangwon National University, Chun-Cheon 200-701 Republic of Korea; *Corresponding author: hongsi@kangwon.ac.kr
INTRODUCTION
When producing structural materials from small- to medium-diameter timber, the processing becomes simpler and more cost-effective than using panels such as glued laminated timber (glulam) or cross-laminated timber (CLT), as the yield of lumber is higher. However, small- to medium-diameter timber may have several defects, such as juvenile wood and knots, significantly reducing its strength and making it disadvantageous for individual use. Laminated timber, which involves gluing many layers of timber, enables the production of large-diameter timber and addresses the structural flaws seen in small- to medium-diameter timber mentioned earlier. Additionally, it is environmentally friendly, as it requires less adhesive due to its smaller bonding surface. Moreover, it enables high-value-added harvesting of thinned logs. Laminated timber has good durability and adhesive performance through relevant research on development projects (Lee et al. 2019). Its potential as a structural component resistant to bending and compression loads has been confirmed (Lee et al. 2021; Baek et al. 2023).
When wide open spaces and longer spans are required in structural elements, moment-resisting structures, such as portal frames, are often necessary to withstand lateral loads (Chen et al. 2018). Modern large-diameter timber construction components are typically connected using metal connectors, particularly through metal dowel joints with slotted steel plates, which provide stability and resistance to moments. Due to the low tensile strength in the fiber perpendicular direction and the low shear strength in the fiber parallel direction, the current design methodology generally ignores the moment resistance capacity and focuses only on transmitting shear force and axial force (He and Liu 2015; Liu and Xiong 2018). Despite this, vertical loads or lateral loads such as seismic loads applied to the beam actually generate moments in these timber column-beam joints. On the other hand, the resistance mechanism against the force exerted on the metal dowel joint with slotted plates can be divided into two types: tensile type, which resists the force applied along the dowel axis, and shear type, which resists the force applied perpendicular to the dowel axis. However, the low tensile strength and shear strength resulting from the fiber direction of the mentioned wood lead to vulnerability due to the wood embedment of the dowel and cause severe destruction of the structure in the direction of end distance, such as cleavage and plug shear (Sawata 2015). These failure modes can cause early failure in ductile structures affected by susceptible forces induced by earthquakes, as they degrade the ductility and energy dissipation capacity. Therefore, enhancing the resistance behavior against lateral loads to ensure the stability of the structure is crucial.
Reinforcing structural elements and joints using fiber reinforced plastic (FRP) has been commonly employed in timber construction (Raftery and Harte 2011; Sena-Cruz et al. 2012; Raftery and Whelan 2014; Schober et al. 2015; Zhu et al. 2017; Grunwald et al. 2018; Toumpanaki et al. 2022). This reinforcement method can improve the flexural performance of the absence and the mechanical characteristics of the joint, particularly demonstrating excellent ability to suppress splitting caused by the fiber direction of wood (Song et al. 2017; Lee et al. 2021; Kim et al. 2022). Above all, FRP possesses several advantageous characteristics that make it suitable for application in timber engineering structures, such as corrosion resistance, high strength-to-weight ratio, and excellent adhesion to wood. There has been a relative lack of research on joints using FRP in structural engineering. The main focus of research has been placed on the ‘glued-in’ joints using FRP rods to replace metal dowel joints. Some researchers have proposed reinforcement methods that use slotted steel plates and metal dowels to improve the moment resistance performance of column-beam joints.
Santos et al. (2013) studied adhesive reinforcement techniques for metal dowel wood joints. The proposed method of reinforcing the area around the hole with carbon fiber reinforced plastic (CFRP) adhesive was found to be more effective in enhancing joint performance and mitigating brittle failure mode (Santos et al. 2013). Kim et al. (2013) examined the shear performance of bolt joints in laminated timber, considering the arrangement of reinforced glass fiber cloth (GFC) and the volume ratio of reinforcement. The research findings indicate that when the glass fiber cloth (GFC) was positioned in the inner layer of the laminated timber rather than the outer layer, the reinforcement effect was more pronounced for the 1% volume ratio reinforced laminated timber. Additionally, for the 2% volume ratio reinforced laminated timber, the yield shear strength was improved by 22% compared to the unreinforced joint (Kim et al. 2013). Xiong et al. (2017) conducted a cyclic load test on a single-story portal frame to investigate the lateral resistance of reinforced glued laminated timber column-beam structures. The results revealed that the reinforcement method of wrapping FRP in the joints could inhibit crack initiation and improve the strength, stiffness, and energy dissipation capacity, indicating that the method using FRP can be applied in engineering practice (Xiong et al. 2017). Yang et al. (2020) researched to enhance the load transfer capacity of dowel-type joints by wrapping FRP straps around single-bolt timber joints externally or attaching them at various angles. Consequently, the strength of the joint was increased by 50% compared to the control group joint, depending on the appropriate fiber orientation of the FRP. Additionally, the FRP prevented early splitting of the joint, mitigating the failure of the timber. The authors conducted studies to strengthen the timber column-beam joint by FRP reinforcement. The aim was to replace metal dowels and steel plates by reinforcing the plywood with fiberglass and laminating it at high temperatures, as described in Song et al. 2014(a, b). Additionally, the structural usability of the laminated beam-column joint was evaluated by inserting the laminate into the joint slot and bonding it, as assessed by Jung et al. 2016(a, b). Previous studies imply that the optimal method for enhancing the performance and constructability of a metal dowel joint is to attach FRP laminates with fiber orientation orthogonal to the joint on the outer side of the joint. If a wooden panel is attached to the outer surface of this FRP laminate, it would be possible to improve the aesthetics lacking in previous studies by attaching or covering the FRP on the outer side of the joint.
This paper investigated the behavior of a laminated timber portal frame reinforced with FRP laminates and a slotted-in steel plate drift joint under cyclic lateral loads of low frequencies. The FRP laminates were produced in three types: GFC in fabric form, GFRP sheet in plastic form, and CFRP sheet. The performance enhancement of laminated timber portal frame joints using these laminates was compared to portal frames without reinforcement. The hysteresis loop was used to evaluate the rotational stiffness, moment resistance capacity of the joint, and its seismic behavior under cyclic load. Additionally, the maximum capacity and ductility of the structure, including the failure mode, were assessed by the monotonic load to failure after the cyclic load.
EXPERIMENTAL
Material Properties
Laminated timber
Lumber refers to the dried timber obtained from larch (Larix kaempferi (Lamb.) Carr.) with a medium diameter, including the pith. The cross-sectional dimensions of the lumber were 120 mm (height) × 90 mm (width). The average air-dry moisture content of the lumber was 15.0 ± 0.9%, and the average air-dry specific gravity was 0.52 ± 0.03. The laminated timber comprised four lumber pieces with an elastic modulus of 10 GPa. The timber was pressed with a pressure of 1.0 MPa after being coated with a phenol-resorcinol formaldehyde adhesive at a level of 400g/m2 (single spread) in both flatwise and edgewise orientations (Lee et al. 2019). The laminated timber, which was successfully produced, was ultimately processed into a cross-sectional size of 220 mm (height) × 170 mm (width) and used for fabricating portal frames.
FRP laminates
FRP laminates were created by attaching FRP reinforcements to lumber boards to strengthen the portal frame joint made of laminated timber. FRP laminates were manufactured in three different types: glass fiber cloth (GFC) woven with fiberglass filaments crossing each other at right angles, glass fiber reinforced plastic (GFRP) sheets pultruded with resin and longitudinally oriented fibers, and carbon fiber reinforced plastic (CFRP) sheets (Fig. 1). The reinforcement materials were stacked with a reinforcement ratio of 6.7% (thickness: 6 mm) based on the cross-section of the joint in the portal frame. Kim and Hong (2015) evaluated the fracture toughness of fiberglass laminated timber and reported that when the crack propagation direction of the wood aligns with the fiber array direction of the reinforcement, the fracture toughness of the laminated timber is significantly reduced. According to previous research findings, it has been observed that the two reinforcing materials, excluding the fabric-type GFC, can easily split along the fiber direction of the reinforcement due to embedment by dowels at the joint. To prevent this, GFRP and CFRP were laminated with their fiber orientations perpendicular to each other. The reinforcing materials were manufactured by applying epoxy adhesive and hot pressing them at a temperature of 60 ℃ and a pressure of 1.0 MPa. After curing, they were bonded to a lumber board with a thickness of 18 mm using Polyurethane (PUR) adhesive (P84, Otto-Chemie, Bavaria, Germany). The final dimensions of the FRP laminates were 24 mm × 220 mm × 240 mm (Fig. 2).
Fig. 1. The manufacturing process of FRP laminates
Fig. 2. FRP laminates bonded with lumber board
Test Methods
Fabrication of laminated timber portal frames
The portal frame was constructed with a single span on the ground floor, with a height of 1690 mm from the floor to the center of the beam and a distance of 1930 mm between the centers of the columns (total height: 1800 mm, total width: 2370 mm). The joints of the portal frame were fixed using an inserted drift pin method, as shown in Fig. 3 (a and c). Slots with a thickness of 10 mm and a depth of 240 mm were machined on the column and beam joints. One rolled steel plate, according to KS D 3503 standards (2018), with a thickness of 8 mm, a width of 220 mm, and a length of 440 mm was inserted. The plate was then attached to the beam and column using 16 drift pins with a diameter of 16 mm and a length of 170 mm. The hole positions of the drift pin were designed to have an end distance, edge distance, and spacing, all measuring 3.5 times the diameter of the drift pin (Fig. 3 (b and d)). In the case of the reinforced portal frame, the columns and beams of the joints were partially reinforced with FRP laminates on the outer side, with a depth of 24 mm, width of 220 mm, and length of 240 mm each (Fig. 3 (c and d)). The wooden part of the FRP laminates was attached to the outside side of the FRP surface by applying PUR adhesive. Four types of portal frames were produced, with two frames per type: unreinforced portal frame (UR), portal frame reinforced with GFC laminate (GFC), portal frame reinforced with GFRP laminates (GS), and portal frame reinforced with CFRP laminates (CS) (Table 1).
Fig. 3. Layouts of the laminated timber portal frame. (a), (b) Configuration of the laminated timber portal frame. (c), (d) Configuration of the laminated timber portal frame with FRP laminates.
Table 1. Geometry and Joints of Laminated Timber Portal Frames
Test setup and instrumentation
Each portal frame, as indicated in Fig. 4, was fixed to the foundation with two steel angle brackets and four high tensile bolts (diameter: mm) per column. Five displacement sensors were installed to obtain the drift angle of the joint. The holder connected to the hydraulic jack was fixed to the beam to transmit the lateral load to the portal frame. To measure the horizontal displacement of the portal frame, displacement gauges (, ,) were symmetrically placed at the same height on both sides of the hydraulic jack’s plate slot opposite the joint (Fig. 5). Furthermore, a single column () was installed at a height of 210 mm from the foundation base. For the measurement of vertical displacement of the portal frame, two columns (, ) were installed at the bottom of the foundation, each at a height of 100 mm from the ground.
Fig. 4. Test setup for laminated timber portal frames
Fig. 5. Test setup and instrumentation for laminated timber portal frame
Loading scheme
The cyclic load tests on the portal frame were conducted using the displacement-centered method. The cyclic load schedule assumes that no vertical load was applied to the structure, according to the Architectural Institute of Japan’s (2012) standards. The portal frames were subjected to three cyclic loading cycles in positive and negative directions within each drift angle section (1/450 rad., 1/300 rad., 1/200 rad., 1/150 rad., 1/100 rad., 1/75 rad., 1/50 rad.), as depicted in Fig. 6. The test was terminated if the joint fractured or if the drift angle exceeded 1/15 radians while moving forward. The velocity of the load was 40 mm/min.
Fig. 6. Cyclic loading protocol
Table 2 presents the horizontal displacement transformation values for the drift angle and the information on the frequency. The joint’s effective drift angle (θ) was calculated using the following Eqs. 1, 2, and 3,
where θ represents the effective drift angle of the beam with respect to the plate, is the drift angle of the beam with respect to the column, and signifies the drift angle of the plate with respect to the column. Furthermore, represent the displacements measured by LVDT 1, LVDT 2, LVDT 3, LVDT 4, and LVDT 5, respectively.
Table 2. Horizontal Displacement According to Drift Angle
RESULTS AND DISCUSSION
Hysteresis Loop
The hysteresis loop serves as a fundamental basis for seismic design, illustrating the relationship between the moment and rotation of portal frame joints under low-cycle loading conditions. The hysteresis loop for each series is graphed in Fig. 7. The moment on the vertical axis represents the total load applied to the timber portal frame joints, which directly cause the stresses generated in the timber portal frame joints. The drift angle on the horizontal axis is the angle of deformation between the column and beam caused by the moment. The hysteresis loop of the portal frames demonstrated no significant damage during the process, namely in the final drift angle section of 1/50 rad. However, it was observed that the reinforcement in the portal frame joint and the type of FRP laminates used resulted in variations of the slope and magnitude of the hysteresis loop. These differences are speculated to arise from the capacity and stiffness of the joint’s load cycle during the first drift angle, as well as the strain hardening of the drift pin and timber during the load cycle, and the strain inhibition performance of FRP laminates based on the reinforcement material. The average maximum moment obtained from the hysteresis loop was 11.6 kN·m, which was the lowest among the portal frames. The GFC laminate, based on the UR, increased the moment of the portal frame by 12%. The moments of GS and CS increased by 22% and 21%, respectively, implying that the laminate manufactured with fabric-type reinforcement had a higher improvement effect on the moment in the cyclic load section than the laminate manufactured with fabric-type reinforcement.
Fig. 7. Hysteresis loops of portal frames
Strength Degradation
The hysteresis loop for cyclic load on structures with joints decreased the load due to the gaps between the joint and the component when compared to the initial load within the same deformation section. The occurrence of defects was determined by the processing form of the joint and the material properties of the joined materials. Particularly, timber structures joined with metal fasteners are susceptible to decreased joint strength between timber members during cyclic loading. The decrease in intensity is ultimately a crucial indicator of the ability to withstand repetitive seismic movements (Pozza et al. 2014; Sheikhtabaghi 2015). Strength degradation () refers to the decrease in load when a constant displacement is applied. The value of was derived from the difference between the maximum load () at the first-order force and the maximum load () at the third-order force within the same positive deformation section based on the hysteresis loop of the portal Rahmen frames in Fig. 7, as described in Eq. 4.
(4)
Figure 8 represents the average strength degradation of portal frames according to the drift angle. No observed correlation was found between the increase in drift angle and the degradation of strength in the case of UR. The strength degradation of the unreinforced portal frame did not exhibit a specific trend and occurs randomly in all drift angle sections. The highest strength degradation of 4.8% was observed in the last drift angle section (1/50 rad.). However, portal frames reinforced with FRP laminates generally did not exhibit strength degradation in the initial drift angle section. However, as the drift angle increased, there was a gradual increase in strength degradation. In particular, GS and CS improved their strength degradation in the low drift angle section.
In the low drift angle section, UR decreased its strength by nearly 4%. However, across the entire drift angle section, there was a non-directional decrease in strength influenced by the hysteresis phenomenon of the joint due to the heterogeneity of the timber part. In contrast, the reinforced portal frames did not decrease in strength in the initial drift angle section but gradually decreased in strength as the drift angle increased. The decrease in their strength did not exceed 2%, and, unlike UR, the phenomenon of joint history was controlled by FRP-reinforced laminate, resulting in minimal variation in strength degradation.
Fig. 8. Average strength degradation at each drift angle
Skeleton Curve
Figure 9 depicts the portal frame’s skeleton curve, which is an envelope curve obtained by connecting the peak points of each first cycle of the hysteresis loop in Fig. 7. The stiffness of the portal frame joint recovered from the previous drift angle stage may be determined by using the skeleton curve, and the tendency of the developed curve can be used to assess the potential for early failure of the structural system.
The portal frames were within the elastic range for the applied lateral load, ranging from 1/450 rad. to 1/75 rad. in the cyclic load section. The reinforcement of portal frame joints by FRP laminates impacted the initial stiffness. The reinforced portal frame exhibited a relatively higher stiffness than the unreinforced portal frame in the first drift angle section. The initial stiffness enhancement of the portal frame under cyclic loading allowed for the continuous maintenance of joint stiffness even with increased deformation, which was observed in dead load and live load conditions. The reinforced portal frame, which is reinforced with fabric-type FRP, slightly decreased its stiffness in the 1/50 rad section for positive and negative loads. Moreover, it was observed that its stiffness recovery is slightly lower than other FRP laminates in a larger drift angle section.
Fig. 9. Average strength degradation at each drift angle (UR: Unreinforced, GFC: Glass fiber cloth, GS: Glass fiber reinforced plastic sheet, CS: Carbon fiber reinforced plastic sheet)
Maximum Load and Failure Mode
Failure mode
The maximum moment of the portal frame was determined by positive loading after applying a cyclic load from 1/450 rad. to 1/50 rad. until the drift angle reached 1/15 rad. or the specimen failed. The yield strength was determined using Yasumura’s proposed method, and the stiffness (K) was defined as the slope of the secant line at 10% and 40% of the maximum load (Yasumura1998). The ductility (μ) of the joint was determined by the ratio of the maximum strain () to the yield strain (∆yield). The average maximum moment () of the UR was recorded at 42.5 kN·m, making it the lowest resistance among the portal frames. The portal frames experienced premature failure due to longitudinal shear stress in the timber, originating from the drift pin holes on the outer column or beam, just before reaching the maximum moment during the last monotonic test after cyclic loading. The uncontrolled premature failure acted as a cause for the low measured yield strength () of the UR. The portal frame exhibited significant fiber direction cleavage from the drift pin holes of all columns and beams, with embedded deformation observed in the outer drift pin holes of the joints (Fig.10). This is a typical failure mode of a wooden structure joint that utilizes a metal dowel. Most drift pins observed after disassembly remained in a straight line without any significant bending deformation.
The reinforced portal frames GFC, GS, and CS showed an average maximum moment increase of 12%, 16%, and 27%, respectively, compared to UR. The average yield moment increased by 38%, 16%, and 10% for GFC, GS, and CS, respectively. The reinforced portal frames did not experience premature failure, unlike UR. Instead, they failed at the maximum moment or after reaching the maximum moment (Fig. 11, Fig. 12, Fig. 13). However, similar to UR, the splitting caused by longitudinal shear stress in the wood still determined the maximum moment. Failure was not observed in the FRP-reinforced laminate on the outer side of the joint, but splitting occurred in the inner wooden part of the joint, similar to UR. An important observation is that plastic hinges were observed in more drift pins than in UR, indicating a higher likelihood of more damage to the wood caused by the drift pins. However, the splitting of timber was significantly reduced in the joints of reinforced portal frames compared to that of unreinforced frames. This confirmed the effectiveness of FRP laminate reinforcement in improving failure mode.
Although no significant difference in the occurrence of cleavage was seen depending on the kind of FRP laminates, the FRP properties caused variations in the deformation of drift pins. Specifically, a greater degree of deflection in the drift spindles of the CS was observed (Fig. 13), implying that the mechanical performance of CFRP laminates was superior to that of GFC or GFRP, and it may be considered the factor that had the most influence on the increase of the maximum moment in a portal frame.
Table 3. Structural Characteristics of Portal Frames with Cyclic Lateral Loads
Fig. 10. Failure modes of laminated timber frame UR
Fig. 11. Failure modes of laminated timber frame GFC
The drift pins of the reinforced portal frame experienced minimal flexural distortion; however, embedded deformation occurred quite significantly in the internal drift pin holes of the laminate. These failure modes were caused by an increase in yield moment () and yield strain (∆yield), and the increase in both strength properties ultimately resulted in decreased ductility. Unlike the skeleton curve for the hysteresis loop, the little difference in stiffness between portal frames in the final yield stage indicates that FRP laminates had a greater impact on cyclic load than monotonic load.